Thinking Machines, Thinking Lawyers
AI and the Future of Legal Research and Knowledge Production - Introduction to the Symposium
This symposium explores the relationship between AI and the production and delivery of legal knowledge. It also interrogates how emerging AI-based capabilities are shaping the way lawyers think, research, and act. Traditionally, lawyering has been shaped through skills acquired by law students and young lawyers on the ground, be it the classroom, social interactions, or the courtroom. Lawyers spend years training how to read complex judgments, parse through jurisprudence, draft legal prose, and finesse their arguments, with crucial sociological insights from their cultural milieu. These skills are acquired through iteration, constant observations, specialized supervision, and most importantly, the space to fail and make mistakes. As the state of technology stands today, many of these tasks could easily be outsourced to companion chatbots and AI-managed workflows. The result is often a near-perfect first draft, a fully formed legal argument tailored to the judicial forum, and frictionless research. This has enormous cognitive and creative stakes for lawyering, and the stability of the legal profession as a whole. There are already reports of mass layoffs and restructuring within law firms to “replace” junior lawyers.
Lawyering in the age of AI is also riddled with contradictions, often between utilitarian and deontological views. For instance, a substantially AI-enhanced brief may increase a lawyer’s success in the courtroom. At the same time, it raises ethical concerns about the lawyer’s claims of ownership over the intellectual product. Consequently, what are the standards of permissibility when using AI to proofread a draft versus delegating the writing of the draft to AI itself? Similarly, does efficiency in terms of shaving off time to prepare a legal brief through AI replace the legal prose and persuasiveness of a human-authored one? These inconsistent observations cannot be resolved overnight, but a timely inquiry can help the legal profession understand what is at stake, and what can be mitigated.
Similarly, generative AI poses several crucial questions for legal academics, both philosophical and those concerned with the quality of the product produced by such platforms. There is no singular way of undertaking legal research, and given such methodological variety, what is the utility of AI among such different methods? Are there certain aspects of legal research and writing that can be delegated to AI, and some that must not ever be delegated? Is the AI-generated product worthy of engagement? What impact does employing AI assistance have on the individual and their cognitive capabilities? Answering many of these questions inevitably pushes us towards the most fundamental one: what exactly is the purpose of legal research, and where does the individual, especially one’s experiences, observations, voice, and thoughts, lie in this enterprise? These are extremely hard questions to answer, particularly once we move beyond the unhelpful approach of simply dismissing AI’s relevance. It is important in these discussions to note that while AI platforms may not present themselves as perfect legal researchers, given the pace at which technological change is progressing, it would become increasingly difficult to dismiss AI’s capabilities by simply quoting its substandard outcomes. The conversation needs to be cognisant of both present capabilities and future possibilities.
We are on the cusp of a change, nothing short of a revolution, in terms of how legal education and research are approached, and this symposium is an attempt to initiate a discussion on some of the questions identified above. The symposium pieces delve into the following themes. First, the impact of AI-enhanced lawyering on the ability of lawyers and legal scholars to think and write. This has both positive and negative consequences. AI tools, if used responsibly, can be used to enhance research by parsing through different methodologies (e.g. comparative, archival, socio-legal). However, there is no consensus on the limits of such use. At the same time, the lack of consensus around responsible use has far-reaching consequences for the future of the legal profession as a whole.
Lawyering is a set of tangible hard skills like the ability to understand a judicial decision and its jurisprudential lineage, and argue in a courtroom. There are equally important but invisible soft skills which lawyers and scholars continue to adopt throughout their careers. The ability to interact with clients, empathy, negotiation, cultural sensitivity, and emotional intelligence are some important examples. The rapid delegation of everyday legal tasks to AI not only weakens the hard drafting and research skills, but also eliminates opportunities for lawyers to hone their soft skills. Further, human-authored legal prose is ultimately an articulation of rational legal argument. But what marks it as unique and persuasive is the force of emotion. Lawyers spend years understanding the sophisticated manoeuvres that appeal to judges and clients. Delegating these tasks to technology risks the hollowing out of human creativity, emotions, and professional strategy that undergird seemingly technical prose. The Symposium pieces also explore how AI-enabled “groupthink” affects legal scholarship and threatens to replace the research skills acquired by lawyers and scholars.
Second, the problem of epistemic justice and epistemic pollution. Legal scholarship does not represent everyone fairly. We ask whether AI further entrenches and exacerbates these inequalities. And whether the communities that have historically been affected by the legal structures and courts are going to be further excluded. Inherent biases embedded in AI tools can shape whose knowledge is represented and whose is excluded. There are also numerous instances of AI generating false case citations and hallucinating legal scholarship that does not exist. These sources are regurgitated in judicial pronouncements. Awkward admissions by reputable judges and law firms, post facto, drive home the need to develop sensitivity and awareness around the pitfalls of mindless delegation. Legal education should also develop tools to enable lawyers and judges to identify and combat artificial research and argument.
Third, the aspects around research integrity and accountability. Artificial research and reasoning ultimately affect access to justice. The recent cases of hallucinations and false precedents being cited in court ultimately impact justice delivery. All participants in the justice system, from lawyers to judges, have been complicit in these isolated but growing incidents. Who is tasked with the responsibility of preserving the integrity of research in such cases, and who is the ultimate arbiter?
Lastly, the questions of institutional and regulatory approaches to generative AI. Here, we need to understand if existing mechanisms governing the bar and bench are aligned with the unprecedented pervasion of AI in the profession as well as in the justice system.
This symposium does not claim to present perfect answers, but is an attempt to initiate conversations on the multifaceted changes posed by AI-assisted capabilities to our traditional ways of doing legal research and teaching. We, therefore, also welcome responses from our readers!



